258 METALLURGY OF IRON AND STEEL. requirements should he worded so that manufacturers would be obliged to put sufficient work on large members to render them of proper structure. There is often a confusion of terms in considering the effect of work as represented by a large percentage of reduction from the ingot, and the effect of finishing at a low temperature. This is found most often in the case of plates, for it has been quite a general practice to roll these directly from the ingot in one heat. In order that a piece shall be finished hot enough under this practice, there has been a standing temptation to use a thin ingot; but, on the other hand, it has been almost universally shown that the best results are obtained when a large amount of work is put upon the piece during rolling. SEC. XlVb.—Discussion of Riley's investigations on the effect of work.—The truth of this last statement was disputed by Eiley,* who tabulated the results of testing different thicknesses of plate when rolled from ingots of varying section. In all cases the ingot was either hammered or cogged to a slab and this was reheated before finishing into a plate. His analysis of the records consisted in picking out individual cases and showing that the small ingots gave some results which were equal to those from the large ones, but this method of comparison must be recognized as entirely unworthy of the subject. It is true that the number of tests is very small, and it would not be surprising if the accidental variations in the double working should produce anomalous results; but even taking these very data and making comparisons by the proper system of averages, it will be found that they tell a story exactly opposite from the conclusions formulated by Mr. Riley. In Tables XIV-A and XIV-B such figures are presented. In the comparison of the different thicknesses in Table XIV-A the thinner plates give much better results, the one-half-inch plate showing an increased ductility in spite of its greater strength. The one-quarter-inch plates are somewhat lower in elongation and two and one-half per cent, better in reduction of area than the one inch plates, but they possess 7600 "pounds more strength, so that less ductility should be expected. This statement is open to criticism, as no account is taken of the effect of variation in the * Some Investigations as to the Effects of Different Methods of Treatment of Mild Steel in the Manufacture of Plates. Journal J. and 8. L, Vol. I. 1887. 121.